期刊名称：International Journal of Electronics and Computer Science Engineering
电子版ISSN：2277-1956
出版年度：2012
卷号：1
期号：3
页码：1494-1503
出版社：Buldanshahr : IJECSE
摘要：In this paper, the problem of deblurring and poor image resolution is addressed and extended the task by removing blur and noise in the image to form a restored image as the output. Haar Wavelet transform based Wiener filtering is used in this algorithm. Soft Thresholding and Parallel Coordinate Descent (PCD) iterative shrinkage algorith m are used for removal of noise and deblurring. Sequential Subspace Optimization (SESOP) method or Line search method provides speed-up to this process. Sparse-land model is an emerging and powerful method to describe signals based on the sparsity and redundancy of their representations. Sparse coding is a key principle that underlies wavelet representation of images. Coefficient obtained after removal of noise and blur are not truly Sparse and so Minimum Mean Squared Error estimator (MMSE) estimates the Sparse vectors. S parse representation of signals have drawn considerable interest in recent years. Sparse coding is a key principle that underlies wavelet representation of images. In this paper, we explain the effort of seeking a common wavelet sparse coding of images from same object category leads to an active basis model called Sparse-land model, where the images share the same set of selected wavelet elements, which forms a linear basis for restoring the blurred image
